TMS320F28035-Q1: Codes change => Flash content change => impact ADC sampling result?

Will,

Will Gao said:

But they still want to check with us why a tiny codes change (Flash contents change) will impact the ADC sampling result.

Which flash change are you referring to? 

Best Regards,

Ben Collier

Codes + rodata contents change.  

Will,

So they changed code unrelated to the ADC functionality, but they see a change in all of their ADC results? 

Yes, it is. Please help to analyze this wired phenomenon. Thanks. 

Ok, I do not think they will really be affected by the errata since they are averaging so many samples. 

Will Gao said:

they did firmware update on field products

Are you able to share more about what changed in the flash? Did they change anything related to interrupts? I'm wondering if there is anything that could interrupt this sequence of ADC conversions. If the ADC goes a long period of time without doing a conversion, there is a chance that the next conversion could be affected by the errata. 

Also, just to be 100% sure that this was the only change, have they tried putting the old firmware back onto one of the devices? Do the ADC samples go back to the old value? 

Got it. I will let them check whether the updates impact the interrupt timing sequence. This issue was found in car, so they can't rollback to old version firmware to debug this issue. 

Ok, I am curious to see your response.